GARNER  V.  JONES (99-137) 529 U.S. 244 (2000)
164 F.3d 589, reversed and remanded.
[ Kennedy ]
[ Scalia ]
[ Souter ]
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version

Souter, J., dissenting


No. 99—137



[March 28, 2000]

    Justice Souter, with whom Justice Stevens and Justice Ginsburg join, dissenting.

    I think the Court of Appeals made no error here and so respectfully dissent from the reversal. A change in parole policy violates the Ex Post Facto Clause if it creates a “sufficient,” California Dept. of Corrections v. Morales, 514 U.S. 499, 509 (1995), or substantial risk that the class affected by the change will serve longer sentences as a result.1 To determine the likelihood that the change at issue here will lengthen sentences, we need to look at the terms of the new rule, and then at the possibility that the terms are mitigated by a practice of making exceptions.

    Before the board changed its reconsideration rule, a prisoner would receive a second consideration for parole by year 10, whereas now the second consideration must occur only by year 15; those who would receive a third consideration at year 13 will now have no certain consideration until year 23, and so on. An example of the effect of the longer intervals between mandatory review can be seen by considering the average term served under the old rule. In 1992, a member of the Georgia Legislature stated that the average life-sentenced inmate served 12 years before parole. See Spotts, Sentence and Punishment: Provide for the Imposition of Life Sentence Without Parole, 10 Ga. St. U. L. Rev. 183, 183, and n. 4 (1993). Some prisoners must have been paroled before 12 years. But those who would have been paroled when considered a second time at year 10 or a third time at year 13 will now be delayed to year 15. While the average helps to show the effects Georgia’s new rule is likely to have on some prisoners who would be released at the early end of the parole spectrum, the changed rule threatens to increase punishment for all life-sentenced prisoners, not just those who would have been paroled at or before the average time. If a prisoner who would have been paroled on his fourth consideration in year 16 under the old rule has to wait until his third consideration in year 23 under the new rule, his punishment has been increased regardless of the average.

    Georgia, which controls all of the relevant information, has given us nothing to suggest the contrary. It has given us no basis to isolate any subclass of life prisoners subject to the change who were unlikely to be paroled before some review date at which consideration is guaranteed under the new rule. On the contrary, the terms of the rule adopted by the State define the affected class as the entire class of life-sentenced prisoners, and the natural inference is that the rule affects prisoners throughout the whole class. This is very different from the situation in Morales, in which it was shown that 85% of the affected class were found unsuited for parole upon reconsideration. Morales, supra, at 511. At some point, common sense can lead to an inference of a substantial risk of increased punishment, and it does so here.

    The significance of that conclusion is buttressed by statements by the board and its chairman, available at the board’s official website, indicating that its policies were intended to increase time served in prison. See Georgia State Board of Pardons and Paroles, News Releases, (Jan. 2, 1998, release, Policy Mandates 90% Prison Time For Certain Offenses) (“Since 1991 the Board has steadily and
consistently amended and refined its guidelines and
policies to provide for lengthier prison service for
violent criminals”); Georgia State Board of Pardons and Paroles, Violent-Crime Lifers Who Died in Prison (June 4, 1998) (quoting Chairman Walter Ray as stating that “ ‘obtaining parole on a life-sentence is increasingly rare’ ” and reporting that “[b]ecause of strict sentencing laws as well as the Board’s conservative paroling policy, agency officials predict successive fiscal years will reflect a rising number of inmates for whom a life sentence does indeed mean just that”).2 If respondent had ever been allowed to undertake discovery, further statements of punitive intent may well have been forthcoming. Although we have never decided that a purpose to increase punishment, absent a punitive effect, itself invalidates a retroactive policy change, see Lynce v. Mathis, 519 U.S. 433, 443 (1997), evidence of purpose certainly confirms the inference of substantial risk of longer sentences drawn above. It is, after all, reasonable to expect that members of a parole board acting with a purpose to get tough succeed in doing just that.

    On the other side, there is no indication that the board adopted the new policy merely to obviate useless hearings or save administrative resources, the justification the Court accepted in Morales. See 514 U.S., at 511. Indeed, since a parole board review in Georgia means that one board member examines an inmate’s file without a hearing and makes a decision, and no specific findings are required to deny parole, any interpretation of the rule change as a measure to conserve resources is weak at best, and insufficient to counter the inference of a substantial risk that the prisoners who will get subsequent manda-
tory parole considerations years after the reviews that the old rule would have guaranteed will in fact serve longer sentences.3

    Thus, I believe the Eleventh Circuit properly granted summary judgment for respondent. Although Georgia argues that the board freely makes exceptions to the 8-year rule in appropriate cases, the State provided no evidence that the board’s occasional willingness to reexamine cases sufficiently mitigates the substantial probability of increased punishment. While the majority accepts the argument that, even without evidence of practice, the board’s discretion to revisit its assignment of a reconsideration date may be sufficient standing alone to preclude an ex post facto challenge, this is surely wrong. The policy statement on which the majority is willing to rely, see App. 56, gives a prisoner no assurance that new information or changed circumstances will matter, even assuming that prisoners are aware (and able to take advantage) of their limited ability to ask the board to change its mind. Because in the end the board’s ability to reconsider based on a “change in [a prisoner’s] circumstance or where the Board receives new information,” ibid., is entirely discretionary, free of all standards, an 8-year period before further consideration of parole made solely upon review of an inmate’s file has to create a real risk of longer confinement.

    A further word about the absence of record evidence of practice under the new rule is in order. One reason that there is none is that Georgia resisted discovery. In this Court, it sought to compensate for the absence of favorable evidence by lodging documents recounting parole reconsiderations before the mandatory reconsideration date. But every instance occurred after the Eleventh Circuit had ruled against the State.4 These examples of reconsiderations are the parole equivalent of fixing the broken front steps after the invited guest has slipped, fallen, and seen a lawyer; they do nothing to show that the board’s own interpretation of its policy mitigated the risk of increased punishment.5

    I also dissent from the Court’s failure to require discovery on remand. At the very least, the record gives reason to expect that discovery could show that the affected class has been subjected to the risk of increased sentences. Morales stressed that the question of what changes will be “ ‘of sufficient moment to transgress the constitutional prohibition’ must be a matter of ‘degree,’ ” 514 U.S., at 509 (citation omitted) (emphasis in original). Even if I am wrong and respondent cannot prevail on this record, it is plain that further discovery is justified to determine the degree to which the change at issue here altered sentence lengths.


1. In the first instance, at least, our cases have traditionally evaluated the effect of the change on the class subject to the new rule, rather than focusing solely on the individual challenging the change, Weaver v. Graham, 450 U.S. 24, 33 (1981). It can be difficult, if not impossible, for one person to prove that a change in penal policy has increased the quantum of punishment beyond what he would previously have received, since a sentencing decision is often a mix of rules and discretion. See Lindsey v. Washington, 301 U.S. 397, 401 (1937). At the same time, when one looks at the affected class it can be quite clear that punishment has increased overall. That is proof enough that the new rule applied retroactively violates the Ex Post Facto Clause and, as an invalid rule, should not be applied to anyone within the class.

2. As Georgia’s punitiveness increased, the number of persons on parole decreased. See Georgia State Board of Pardons and Paroles, Georgia’s Criminal Justice Population Increased by 9% in 1998;
Only Decrease Was in Persons on Parole
NewRelea.nsf, (Feb. 1, 1999).

3. The majority suggests, ante, at 7, that the Court required no particular procedural safeguards in California Dept. of Corrections v. Morales, even though the Court mentioned those safeguards as an important factor in its conclusion that there was no increase in the quantum of punishment in that case, see 514 U.S., at 511—512. This is true, but it does not address the problem with Georgia’s virtually unbounded scheme. Once the risk of increased punishment exists, the Board’s nearly nonexistent safeguards provide no way of reducing that risk. Georgia insists that its lack of procedural safeguards is irrelevant to this case, because due process does not require much in the way of procedural safeguards for parole. But that is beside the point. The challenge here is to the retroactive increase in the quantum of punishment. Unlike the California procedure for delaying parole reconsideration in Morales, the Georgia procedure here includes no actual hearing for the prisoner whose reconsideration is delayed five extra years, and the Board is not required to explain itself. Georgia’s procedural minimalism increases the likelihood that prisoners will get rubber-stamp treatment, and decreases the likelihood that the exceptions to the policy on which the majority relies will actually be applied in a way that diminishes the significant probability of increased punishment. Cf. Penson v. Ohio, 488 U.S. 75, 81—82, n. 4 (1988) (stating that a requirement to give written reasons provides an inducement to make careful decisions in cases that might otherwise be summarily ignored); Smith v. Robbins, 528 U.S. __, __ (2000) (slip op., at ___) (Stevens, J., dissenting) (noting that the process of writing out reasons for decision improves the quality of the decision and can reveal error). Parole need not operate under rigidly defined procedures, but if the board decides to make changes retroactive, it must do something to prevent those changes from increasing punishment in violation of the Ex Post Facto Clause.

4. Georgia’s statistics show only that, in fiscal year 1999, about 20% of inmates received reconsideration dates of three years or less; about 10% got reconsideration dates more than three years but less than eight, and 70% got 8-year dates. See App. to Reply Brief for Petitioners 9. Eighty percent were therefore at least potentially negatively affected by the change from a 3-year to an 8-year delay in reconsideration. Even on their own terms, then, the statistics do not show that board policies mitigate the substantial risk of increased punishment.

5. Indeed, as the board explains its decisionmaking procedures, “[t]he overriding factor in determining whether or not to parole a person under life sentence is the severity of the offense.” Georgia Board of Pardons and Paroles, Parole Decisions
Decisions.htm (visited March 2, 2000). If we accept the Board’s statements, changed circumstances or new information would rarely make a difference.